It has been observed also that when t

18 Pages·2008·1.76 MB·English

THE PASSAGE OF FLUID AND PROTEIN THROUGH THE
HUMAN CAPILLARY WALL DURING VENOUS
CONGESTION
By EUGENE M. LANDIS, L. JONAS,1 M. ANGEVINE, AND W. ERB
(From the Robinette Foundation and William Pepper Laboratory of Clinical Medicine,
Hospitalofthe UniversityofPennsylvania, Philadelphia)
(Received forpublication March 22, 1932)
It has been demonstrated in normal human subjects by several
methods that during venous congestion fluid is filtered from the blood
into the tissue spaces. This passage of fluid through the capillary wall
has been identified by comparing blood samples removed during normal
circulation and during venous congestion (Schultz and Wagner (1909);
Rowe (1915); Dautrebande, Davies and Meakins (1923); Peters, Eisen-
man and Bulger (1925)). Ithasbeenobservedalsothatwhenthevenous
pressureinthelegsisincreased byquietstanding, fluidisfiltered fromthe
blood stream with a consequent reduction in circulating blood volume
(Thompson, Thompson and Dailey (1928); Waterfield (1931)). The
filtration produced by measured grades of venous congestion has been
measured also by plethysmographic methods in which the accumulation
of fluid in the tissue spaces was estimated by measuring the increase in
limbvolume (DruryandJones (1927); Krogh,LandisandTurner (1932)).
Thecapillarywallhasusuallybeenregardedasrelativelyimpermeable
to proteins (Krogh (1929); Thompson, Thompson and Dailey (1928);
Krogh, Landis and Turner (1932)). Waterfield (1931) however, differed
from Thompson, Thompson and Dailey in finding that during standing
the blood lost not only fluid but also a significant amount of protein.
Drinker and his co-workers (1931) have called attention to the high con-
centration of protein often found in lymph. Maintaining that tissue
fluidandlymphareidentical, theyregardthecapillarywallaseverywhere
quite permeable to protein even under normal conditions.
A retrograde movement of protein from the tissue spaces through
the capillary wall into the blood during venous stasis has been described
by Plass and Rourke (1927). They found thatduringvenous congestion
the blood proteins were always increased by a greater percentage than
was cell volume. This was believed to indicate that during venous con-
gestion proteins must pass from the tissue spaces into the blood stream.
It will be shown below that the direct comparison of percentage increase
in cell volume with the percentage increase in plasma proteins is unjusti-
Woodward Fellow in Physiological Chemistry.
717
718 CAPILLARY WALL DURING CONGESTION
fied. Actually their figures show loss of both fluid and protein from the
blood stream.
The amount of tissue fluid accumulating in the arm during mild
venous congestion (Krogh, Landis and Turner (1932)) seemed great
enough to produce measurable changes in the fluid content of the blood.
Assuming that the fluid is lost chiefly from the plasma, the amount of
capillary filtrate per 100 cc. of blood can be calculated from the cell vol-
ume of normal blood and the apparent increase in the cell volume of the
blood after passage through the congested vessels.
If the capillary wall were totally impermeable to colloids the protein
content of the blood would be increased in proportion to the reduction
in plasma volume. However, in computations involving changes in
cell volume and plasma proteins, allowance must be made for the fact
thatthecell volume is measured in volumes percentofblood (i.e. plasma
plus cells) while blood proteins are measured in terms of percentage of
plasma. Knowing the change in cell volume produced by stasis and the
protein percentage ofnormal blood, itis possible to calculate the increase
in plasma protein to be expected if no protein were lost in the capillary
filtrate. The difference between the calculated amount of protein and
the amount determined by analysis indicates the amount of protein
lost in the capillary filtrate.
Fromarecalculation ofthedatareported by Plassand Rourke (1927),
and from our own findings, it appears that high grades of venous conges-
tion caused the filtration of relatively large amounts of fluid containing
a high percentage of the three protein fractions present in blood plasma.
Verylowcongestion pressureswereaccompanied bythe filtration ofsmall
amounts of fluid without detectable loss of protein, but the volume
changes were probably too small to permit any certain conclusion con-
cerning protein. Finally, in two patients edema fluid largely produced
by, and collected during, venous stasis contained 0.39 and 0.09 per cent
protein.
METHODS
The blood studies were made in normal male subjects or in male patients
hospitalized for minor maladies unrelated to the circulatory system. Their
blood pressures were not in any instance significantly abnormal; their ages
ranged from 24 to 40 years.
In order to avoid postural changes in fluid balance the subjects reclined
30 minutes before the venous congestion wasstarted. Theforearms were sup-
ported on sandbags at the side of the body, so that the upper (flexor) surfaces
werelevel with theclavicles. Theskin temperatureofeachforearm was meas-
ured at five-minute intervals by means of thermal junctions held, by a single
layer of surgeon's plaster, in contact with the skin on the flexor surface of the
forearm midway between wrist and elbow. Venous pressure was elevated by
means of Riva Rocci armlets, 12 cm. wide and 50 cm. long, wrapped around
the arm well above the elbow.
At the end of thethirty minute rest period thearmletswere inflatedsimul-
taneously and suddenly from reservoirs connected with manometers. On the
E. M. LANDIS, L. JONAS, M. ANGEVINE, AND W. ERB 719
control side the armlet pressure was 9 mm. Hg, a pressure which, though
enough to distend the veins slightly, did not produce measurable filtration in
plethysmographicobservations (Krogh, LandisandTurner(1932)). Thepres-
sure in the other armlet was raised to 20, 40, 60, or 80 mm. Hg in the experi-
mental observations and to9 mm. Hgin threecontrol observations. The sub-
jects were cautioned against moving the arm or forearm during the congestion
period.
When venous pressure had been elevated for a period of 30 minutes sam-
ples of venous blood were removed simultaneously from each arm, approxi-
mately 30 cc. of blood being collected in a syringe containing 20 mgm. of dry
heparin. The armlets were deflated only after the blood samples were re-
moved. The blood was agitated in the syringes for fifteen minutes to dissolve
the heparin and to mix it thoroughly with the blood. The samples were then
transferred to test tubes in which the gentle agitation was continued while
small amounts were removed for determinations of cell volume, hemoglobin
and erythrocyte counts.
Hematocrit determinations were carried out immediately, usually in tripli-
cate, using tubes having an inside diameter of 2 mm. and a length of 12 cm.
A certain number of determinations were also made with Wintrobe hematocrit
tubes. The filled tubes were rotated for 30 minutes at a speed of 3000 r.p.m.
PROTOCOL 1
Venouscongestion of80mm. Hg.
SubjectL. December8, 1931. Roomtemperature 24.20 C.
Armtemperature
Time Apierrmtteumre- Notes
Right Left
p.m. °C. °C. °C.
6:40 Subject reclined, thermal junctions applied and
armletsplaced in position
7:10 33.7 33.4 24.1 Venouscongestion begun
7:15 33.9 33.4 24.3 80 mm. Hgon leftarm
7:20 33.9 33.1 24.2 9 mm. Hgon right arm
7:25 33.9 32.7 24.2
7:30 33.8 32.6 24.2
7:35 33.7 32.3 24.3 Thermaljunctionsremoved
7:40 Bloodsampleof30cc.removedfromeacharminto
syringecontaining 20 mgm. of heparin
Bloodstudies
Congestion Erythrocytes Cell Total Globulin Albumin Nonprotein
pressure volume protein nitrogen
grams grams grams grams
mm.Hg perc.mm. Percent per per per per
100cc. 100cc. 100ce. 100cc.
9........................ 5,596,000* 53.1t 6.58 2.43 4.15 .035
Observed, 80 6,334,000* 60.2t 7.81 3.02 4.79 .034
..........
Calculated, 80 .. 8.71 3.32 5.49
*Average of 8 counts.
tAverage of 3 tubes.
720 CAPILLARY WALL DURING CONGESTION
In certain observations 3 to 5 erythrocyte counts were made on each sample by
each of two observers, the results being averaged to determine the percentage
increase in erythrocytes produced by stasis. Hemoglobin determinations were
made according to the method of Dreyer, Bazett and Pearce (1920). Of the
blood 0.1 cc. was diluted with 19.9 cc. of 0.9 per cent sodium chloride solution.
This suspension of cells was hemolyzed with saponin and the depth of color
in the two specimens was compared by colorimeter, using diffuse daylight..
The blood from the control sidewas takenas 100 percentand only the relative
increase was determined. Plasma proteins were estimated according to the
method of Howe (1921). When the changes in total protein were conspicu-
ouslylarge, the plasma proteins wereseparatedinto globulin and albumin frac-
tions, the former containing also fibrinogen.
Protocols 1 and 2 show the details of experiments at congestion pressures
of 80 mm. Hg and 40 mm. Hg respectively.
PROTOCOL 2
Venouscongestion of40mm. Hg.
Subject R. December 29, 1931. Roomtemperature24.5° C.
Armtemperature
Time pAeirratteumre- Notes
Right Left
P.m. oC. °C. C.
6:39 Subject reclined, thermal junctions applied and
armlets placed in position
7:09 34.4 33.7 24.8 Venouscongestion begun
7:14 34.4 33.0 24.8 40 mm. Hgon left arm
7:19 34.3 33.0 24.2 9 mm. Hgon right arm
7:24 34.3 32.9 24.3
7:29 34.2 32.9 24.3
7:34 34.1 32.7 24.3 Thermaljunctionsremoved
7:39 Bloodsampleof30cc.removedfromeacharminto
syringe containing 20 mgm. of heparin
Bloodstudies
Congestion Hemoglobin Cellvolume Total Nonprotein
preasure clvoue ptin nitrogen
mm.Hg percentinitial percent gramsper100cc. gramsper100cc.
9. 100.0 38.8* 6.17 .028
Observe.d.,..4...0............. 106.3 41.0* 6.80 .026
Calculated, 40 6.79
*Average of two tubes.
METHODS OF CALCULATION
To calculate the loss of fluid per 100 cc. of blood from hematocrit deter-
minationsrequirestheassumption thatthefluidislost entirelyfromtheplasma,
and that in the concentration of the blood the absolute size of the cellular
elements has not changed significantly. Dautrebande, Davies and Meakins
(1923) describe a single observation in which stasis increased the hematocrit
reading by a smaller percentage than the oxygen capacity and the hemoglobin
E. M. LANDIS, L. JONAS, M. ANGEVINE, AND W. ERB 721
content. They calculated that 20 per cent of the fluid lost by the blood must
have been removed from the erythrocytes and 80 per cent from the plasma.
Peters, Eisenman and Bulger (1925), on the contrary, report two observations
in which the application of a tourniquet to the arm for five minutes increased
the oxygen capacity by 20.6 and 23.3 per cent while the cell volume was in-
creased by 19.9 and 25.5 per cent respectively. They conclude that the fluid
lost from the blood during stasis is removed almost entirely from the plasma.
A similar finding is reported by Peters (1924).
Van Slyke, Wu and McLean (1923) observed that as blood changes from
the arterial to the venous state there is a slight but definite increase in cell
volume. In venous congestion the increase in CO2 tension of the blood tends
to increase the cell volume which must, of course, affect the hematocrit read-
ings. The effect is to some extent counteracted by the increased oxygen un-
saturation. This change, though present, is small, amounting to only 0.6 vol-
umes per cent for a change of 30 mm. in CO2 tension even when oxygen-satu-
ration is unchanged (Eisenman, Bulger and Peters (1926)). It did not seem
CO2
necessary nor advisable to collect the blood under oil since the loss of
would only diminish any change of cell volume that had resulted from the
accumulation of CO2 which occurred during stasis.
Duringvenous congestion theconcentration ofthe plasma proteinincreases.
If the wall of the erythrocytes were impermeable to protein alone this change
might be expected to diminish cell volume. Van Slyke, Wu and McLean,
however, on the assumption that the membrane of the erythrocyte is imper-
meable to proteins and to the inorganic cations, regarded the osmotic effects of
the plasma proteins as a negligible factor in the movement of water between
the erythrocytes and plasma because the effective osmotic pressure of the
plasma can be modified very little by change in the plasma protein percentage.
Sudden variations in the number of red cells per unit volume of blood have
been described by Lamson, Abt,Oosthuisen and Rosenthal (1923). Since the
arterial blood entering both arms must have the same composition, systemic
variations in red cell number would be present in both arms equally. To
avoid the effects of these changes blood samples were removed from the contro.
and the experimental arms simultaneously and at approximately the same ratel
In this way each experiment included its own control. The possible effects of
systemic variations in red cell number on the comparison of unconcentrated
and concentrated blood were thereby avoided.
In all the observations here reported the hematocrit readings were matched
with either red cell counts or with hemoglobin determinations to identify any
significant error which might arise from changes in absolute cell volume. No
consistent differences were found (Table 1) between the two series of figures.
Under the conditions of these observations, therefore, it seemed justifiable to
calculate water loss from the changes in hematocrit readings, assuming that
there was no change in the water content of the erythrocytes. Thus, if 100
cc. of unconcentrated blood from the control arm is compared with concen-
trated blood removed simultaneously from the other arm certain relations will
Cl
hold. If the percentile cell volume of the normal blood is and that of the
concentrated blood C2 while the loss of water or volume from 100 cc. of initially
unconcentrated blood is x, then
x = 100- 100 -.
C2
If the capillary wall has allowed no plasma protein to pass during the process
of concentration of the blood, the absolute amount of plasma protein will be
722 CAPILLARY WALL DURING CONGESTION
the same, but the percentage will be increased in proportion to the concentra-
tion of the plasma fraction of the blood. If the protein percentage of the
normal plasma is Pr, and that of the concentrated plasma Pr2, while PF1 and
P!2representtheplasmavolumeof normal and concentrated bloodrespectively
per 100 cc. of initially unconcentrated blood, then, if no gain or loss of water
by the cells is postulated, P12 = PI, - x, and
Pr2 Prl
-x
By comparing the calculated protein percentage, Pr2, with that actually ob-
served, Pr2', the loss of protein may be detected. The actual amount of pro-
tein lost by the plasma from an initial volume of 100 cc. of blood, APr, can be
estimated from the observed protein percentages.
APr =Pri X -T Pr2' X
100 - 100
Knowing the amount of fluid (x) which left the plasma and the simultaneous
loss of protein (APr) the percentage of protein in the capillary filtrate can be
estimated.
OBSERVATIONS
The results are collected in Table 1. Room temperature was kept
between 22.6 and 25.90 C. to avoid marked cooling of the exposed arms
during the observation. The average arm temperature varied between
35.1 and 31.60. Skin temperature fell gradually when higher grades of
venous congestion were used; this fall was never more than 1.20 C. and
was usually less than 1.00 C.
(A). Theloss offluidfrom the blood duringvenous congestion
With a venous pressure of 80 mm. Hg the relative cell volume was
increased by 13.4 to 24.2 percentwhile the hemoglobin and red cellswere
increased by 13.1 to 22.9 per cent. These changes indicate a loss of fluid
amounting to between 11.9 and 19.5 cc. from 100 cc. of whole blood.
With a venous pressure of 60 mm. Hg relative cell volume was in-
creased by 7.8 to 9.7 per cent while red cells and hemoglobin were in-
creased by 7.3 to 10.8 per cent, indicating that the fluid lost amounted
to between 7.2 and 8.9 cc. from 100 cc. of whole blood.
With a venous pressure of 40 mm. Hg the changes were still smaller
and, due probably to the relatively greater error, more variable. Rela-
tive cell volume increased by 1.9 to 5.7 per cent while the hemoglobin
increased by 0.9 to 6.3 per cent, indicating that the fluid loss amounted
to between 1.9 and 5.6 cc. from 100 cc. of whole blood.
With a venous pressure of 20 mm. Hg there was a measurable change
in two of three experiments. Relative cell volume increased by 1.0 to
2.6 per cent and the hemoglobin by 0.7 and 1.6 per cent. The loss of
fluid amounted to between 0 and 2.3 cc. from 100 cc. of whole blood.
E. M. LANDIS, L. JONAS, M. ANGEVINE, AND W. ERB 723
In three observations in which venous pressure was 9 mm. Hg in
both arms there was no change in the relative cell volume in two experi-
ments, and a small but unimportant change in the third.
Plass and Rourke (1927) have reported a series of observations in
which hematocrit readings were made on blood drawn with and without
venous congestion. Those experiments in which measured grades of
venous congestion were used have been collected in Table 2 to show fluid
loss for comparison with our own data. Venous congestion of 100 mm.
Hg for 15 to 22 minutes was accompanied by a loss of between 18.9 and
29.3 cc. per 100 cc. of blood. At 90 mm. Hg the loss was distinctly less,
amounting to between 12.5 and 15.4 cc. per 100 cc. of blood. One
observation at a venous pressure of 80 mm. Hg showed a loss of 7.9 cc.
after 26 minutes. The two series of observations indicate that the loss
of fluid increases as the venous pressure rises.
The relationship between fluid loss and venous pressure is shown
in Figure 1. It is apparent that the loss of fluid is not directly propor-
20~~~~~
N..
Bt/50~~
0
0~~~~~~~
u 20~I 40 60 80
Vows &VeV07- mmAt./
FIG. 1. CHART SHOWING RELATIONSHIP BETWEEN VENOUS CONGESTION AND
LOSS OF FLUID FROM THE BLOOD
tional to the level of venous pressure. The reason for this is discussed
below.
724 CAPILLARY WAiLL DURING CONGESTION
C'
4.) tt%.6 8( (9 00 '-4 o t- 0C; m
-I(
-
it1-0 |e C
3ej~te~tf' ) ~
o~0 ~0 ~ (09 0N 000 0- 0 00
i-m _1- Om00 X _ _e' e0 ' 0 0
cI A ;OrIoXI"toI
O
00~~~~~~~~~~~~~~~~~~~~~~~0
o Wn_tot00 0 00 0 00 t'- o . io..eNe
4g )e + +__
40J 4 t 0 i0f 0f) 014 0' 00: '-z 0%
aQ-55~ts~X ~ (9 - '-4~0~0i 4 0 oe-4 tl '-o4
,4 ~*=0nX~~t-~s~~~~~0 %O o ~10 4 q %00 olo°
rz1%
4iJ
08 o
00
'1 0 (9 t- 0 00
20c 1I 0%I# I 6t0: 1I _C 0% %t- I
'O, t 'I1qe-0"0t''I-.f0' 'VOt'-)-) c'E0'1C4 e'00-~0~%~4'100 0'1'4 01
4)I~~~~~~~~~I ~~I~~~I ~~~C ___l1oo
'41 f' " __ ___
~P~aI:%U kNS I e~~+~0%~~~%'eIO>-N~t I-OOe(9ff)S'0 N I0-0'0-0'I0e0%n| IN
et +|
I I- I It lo 1nlo0
t (00
g0I U + I ee 4tI OeN1otIotm eq~~~~~' 1 I%O4IC+
iSXI° N~~~~~~16 16N6 16 1> el; 1>1
..- 1.to -..v- 4.. 0mvOI4"%
001
%O
V--4
E. M. LANDIS, L. JONAS, M. ANGEVINE, AND W. ERB 725
d)a
4;8 °0 0 0
i&to 0 0 0|
0 0 0
I °00 I0 00 0
0
6ij Cij4
.-0
~' S E * *
1C), - - e- 0 0 0o 0"4
oonOt K 9 99...++9 .e._
u-O4, o° 0o 0 o 0 0£ 000
I___ + I O I b O__ O__ _O
I _
-0 o toC eIfn) 0 t-
z~L~ -0 t- _ _ -__ -
. 23~1~§~~% 6 %6 %6X- 16 % 16 16 r1
t-~ n t- 9
ci C.- C> %O 0 ~~~~~~~'~0~0 it'
24 |zE i T eo _ lo. ° ° o ° |
rz4
'0 z-Dt '0 o o CoD 0 CDo 0
ao :t 80 o 08 0 80 0 0 0
m
C)
4-)
Cd
0P, ji~ ----+'1+1+1+
._-
m + + +
co
S C? C1~~~0'4 0% 00' _c); C4I-_S)
bOCOeJ mEt\C\ CS ON 0\ ON 4
-cO0U
08SSCsf~UB0jjl1I0tm041f 1ot0r1jo~jl1v0J0t0~o11o44+t1t1f+'i--04ftlrjj--jomINo o * 0OA
C-WE*I|S°W$ass(0C40)0-HftNi 1- e'1C-N0 1 CeI--'OIINua%10)-'fO d'i4fC1N -Ct1Z I c'0eE* CiSfI4)i 040
0(f)~~~~~~~~~~ff))tf)~~~~~~f)~~~f) ~~~~f)~~~#) tf)Cf) t(f)~~~#~~~~~~~~f))~~~~f)~~~~~#)~~~~f~)~~~~#~) . 4)g
s d
bi _
ct Cd
Kt
K 1K K Q4U
W.03 01% 1 C0-)a t
O>S3
47
726 CAPILLARY WALL DURING CONGESTION
(B) The passage of protein through the capillary wall during venous
congestion
The observations on the influence ofvenous congestion on the plasma
protein percentage are shown in Table 1. The total protein content of
the blood removed from the control arms (venous pressure 9 mm. Hg)
varied between 5.61 and 7.13 per cent. In the same column below each
normal figure is shown the protein content of the blood removed simul-
taneously from the other arm in which venous pressure was elevated to
80, 60, 40, 20, or 9 mm. Hg. For purposes of comparison in the next
column are given figures which represent the concentration of protein
which, if the capillary walls were impermeable to protein, should have
been produced by the fluid loss observed.
In one of the five experiments at 80 mm. Hg pressure the observed
and calculated protein concentrations agreed, indicating that little
or no protein had been lost. In the remaining four experiments the
calculated protein content was greater than the observed protein content
by 0.38 to 1.18 per cent indicating asignificant loss through the capillary
wall.
Knowing the amount ofplasma, the apparent loss of protein from 100
cc. of blood was computed according to the method described above.
Dividing the protein loss (grams) by the amount offiltrate formed simul-
taneously provided a rough estimate of the protein percentage of the
capillary filtrate. With a venous pressure of 80 mm. Hg the filtrate
appeared to contain between 0.1 and 2.8 per cent protein, indicating that
under those conditions the capillary wall is far from impermeable to
protein.
Avenous pressure of 60 mm. Hg produced asignificant loss of protein
in oneexperiment (number 6); the discrepancy between the observed and
calculated protein percentage amounted to 0.14 per cent, corresponding
to 0.7 per cent protein in the 8.9 cc. of filtrate removed from 100 cc. of
blood. In the remaining three experiments at this pressure no really
significant discrepancy existed between the observed and calculated
values. Had 8.0 cc. of capillary filtrate containing even 1 per cent of
protein been filtered frombloodwith aplasmavolume of 50 per cent, the
discrepancy between observed and calculated protein should have been
0.16 per cent. This difference, if actually present, should have been
detected, since the control analyses of protein (Experiments 13, 14, and
15) agreed within 0.02 to 0.06 per cent. The capillary filtrate produced
at a venous pressure of 60 mm. Hg contains, therefore, relatively little
protein, averaging in four experiments 0.3 per cent.
The amount of fluid filtered from the blood by a venous pressure of
40 mm. Hg is small in comparison with the combined errors of the
hematocrit and protein determinations. While no loss of protein was to
be detected, this finding is of no significance since, had the filtrate con-